he idea of the immune system as exquisitely sensitive to individual pathogenic threats is by now deeply ingrained. The classic picture is of antibody genes whose components can be shuffled at random. One combination happens to fit the antigens presented by a particular invader, and the cells that produce it are then selected for and encouraged to multiply. The specific antibody is produced in quantity and targets the invader by attaching to the antigen, so that the rest of the immune system can come along and clean up. Thus do mammals acquire adaptive immunity.

But there is an entirely separate innate immune system which reacts non-specifically to a wide range of pathogens. It protects the host in the earliest phase of an infection, before adaptive immunity comes into play. White blood cells respond to picogram quantities of lipopolysaccharide (LPS), a complex glycolipid found in the cell wall of many Gram-negative bacteria, by secreting cytokines. In small quantities the cytokines activate natural killer cells and the epithelial barriers of the body. They are also vital to adaptive immunity, because they co-stimulate naive T cells, along with the antigens, which helps select the correct antibody. But the output of cytokines has to be finely balanced. Too little, and the pathogen succeeds. But too much leads to fever, lung dysfunction, circulatory collapse, and kidney failure–septic shock.

The innate immune system protects all animals studied to date, from the nematode worm Caenorhabditis elegans to humans. Elements are even present in plants. The previous issue of Science Watch (11 [4]:1-2, July/August 2000) placed a key aspect of innate immunity at the head of a list of cutting-edge research fronts–rapidly evolving fields identified by citation analysis, each built on a foundation of "core" papers. The first occurrence of innate immunity in the What's Hot list–in the form of Toll-like receptors–is currently at #5.

Toll-like receptors (TLRs) are named for the Toll gene of Drosophila, a trans-membrane protein that has two quite distinct functions. During development, Toll guides the dorso-ventral distribution of cells across the body axis. Later, it is a vital component in the fly's immune response to fungal infections, generating signals that promote the secretion of generalized anti-fungal peptides. Tantalizingly, there were hints that Toll might be related to the mammalian innate immune system: Toll shares cytoplasmic sequences with the Interleukin-2 receptor, and IL-2 is a key component of the innate immune response.

In 1997 two groups cloned human versions of Toll which were homologous right along their length. Fernando Bazan's group at the DNAX Research Institute in Palo Alto isolated 5 hTLR genes and mapped them to their chromosomes (see F.L. Rock, et al., PNAS, 95[2]:588-93, 1998). Howard Hughes Medical Institute (HHMI) researcher Charles Janeway and his colleagues at Yale University independently isolated hTLR genes and showed that they can activate the production of cytokines (see R. Medzhitov, et al., Nature, 388:394-7, 1997).

At #5 is another of the core group of 11 papers that defines ISI's research front on Toll-like receptors. Paul Godowski's group at Genentech in San Francisco claimed that TLR2, one of five hTLRs, is "a direct mediator of signalling by LPS." A human kidney cell line was transfected with a copy of TLR2. When stimulated with LPS and LBP (LPS-binding protein)–but not with either LPS or LBP alone–these cells activated TLR2. Mutants of TLR2 that lacked the intracellular portion did not activate the downstream elements of the pathway, while those with a different extracellular portion also failed to transmit the signal.

There is, however, more to the story than that. Also in the core group of TLR papers are publications showing that TLR4–rather than TLR2–is the primary receptor for LPS. In Montreal, a study reported by S.T. Qureshi and colleagues (see J. Exp. Med., 189[4]: 615-25, 1999) looked at two strains of mice that are innately less responsive to LPS and therefore naturally tolerant. Both strains had a mutation that located to the region of the TLR4 gene. In one, there was a deletion of TLR4, in the other a point mutation that substituted a highly conserved proline by histidine. A group under HHMI investigator Bruce Beutler at the University of Texas Southwestern Medical Center, Dallas, independently confirmed these results using a different technique (see A. Poltorak, et al., Science, 282:2085-8, 1998). And Shizuo Akira's group at Hyogo College of Medicine in Japan generated mice that lacked the TLR4 gene (see K. Hoshino, et al., J. Immunol., 162[7]:3749-52, 1999). Although they grew normally (which suggests that, unlike in Drosophila, TLRs in mice are not important for development), these knockout mice proved resistant to challenge with LPS. Akira's group also confirmed the point mutation in one of the tolerant strains of mice.

All those publications are still outside the Top Ten, but if past experience of research fronts is anything to go by we can expect them momentarily. Toll-like receptors, meanwhile, are finding wide application, with the hunt for a treatment of life-threatening sepsis among the most urgent.

Science writer Dr. Jeremy Cherfas
works with the Biotechnology and Biological Sciences
Research Council of the U.K., Swindon.